Some meetings need a room, not a link.
I had an experience this week that I want to talk about, because it surfaced something I thought we had all figured out, and we have not.
I spent time with my team in person this week. Not on a video call. Not on a hybrid setup with half the team on a screen and half in the room. In person. Same physical space. Same coffee. Same lunch. Same hours together. It is not a thing I get to do as often as I would like, given how the team is distributed and how the work moves. But this week I did, and I noticed something I had let myself forget over the last few years of running everything through video.
The room hits differently.
I do not mean that as a vague nostalgia line. I mean it precisely. There were specific moments this week where the value of being in the room could not have happened any other way, and where the absence of being in the room would have cost the company something real.
Here is what I noticed.
What does it feel like to be in the same physical room as your team?
Being in the same room with your team transmits signals that video calls cannot reproduce. The temperature in the room when a difficult topic comes up, the texture of silence after a hard question, the body language that telegraphs hesitation before someone speaks, the difference between an active listener and a passive one. These signals evolved across millions of years of human cooperation and do not travel through a webcam, an AI meeting summary, or any current remote tool.
You feel the room. That phrase sounds soft, but it is doing real work. When a difficult topic comes up, you can feel the temperature in the room shift in a way that no video call platform reproduces. Bodies adjust. Someone leans back. Someone else leans in. The person who is about to push back does a small physical thing that telegraphs the push-back before they speak. The person who has reservations they have not yet articulated holds their breath for a half-second longer than usual. These are signals that have evolved across millions of years of human cooperation, and they are the signals that good leaders have been reading their entire careers. They do not transmit through a webcam. They certainly do not show up in an AI-generated meeting summary.
What does silence in a room tell you that silence on video does not?
In a physical room, the silence after a hard question has shape. You can feel where the discomfort is coming from, who is processing, who has an unspoken concern, and who is quiet because they need to be drawn out by name. Over video, silence is flat, and the texture of the room is lost in the grid.
You feel the hesitation. When you ask a room of people for input on a hard decision, the silence after the question is a piece of data. The texture of that silence tells you whether the room is processing, or whether the room is uncomfortable, or whether the room has a specific concern someone is trying to figure out how to raise. Over video, the silence is flat. It feels like a pause, or a connection lag, or a question of whether someone is going to unmute. In the room, the silence has shape. You can feel where it is coming from. You can look at the people who are quietest and tell whether they are processing well or struggling. You can ask the person whose body language has shifted, by name, and get an answer that you would not have gotten on a call.
How can you tell who is bought in?
In person, the difference between an engaged participant and a half-paying-attention one is immediate and obvious from body language, posture, and the willingness to riff on someone else's idea. Over video, the grid flattens engagement and everyone looks roughly the same, which means leaders learn less about their team in a quarter of video calls than they learn in a single in-person meeting.
You see who is bought in. On video, everyone looks roughly the same. The grid flattens engagement. The active listeners and the half-paying-attention attendees occupy identical real estate on your screen. In the room, the difference is immediate and obvious. The person who is leaning forward and asking questions and writing things down is doing a different kind of work than the person who is nodding politely. The person who interrupts to riff on someone else's idea is signaling alignment that the muted-on-Zoom version of the same person cannot signal. You learn things about your team in a single in-person meeting that you would not learn in a quarter of video calls.
What happens around a meeting that an AI summary cannot capture?
A large portion of the most important conversations at a company happen outside the formal meeting: in the hallway, at the coffee refill, in the parking lot, at dinner after the meeting, in the office after the door closes. None of those are captured by AI note-takers or meeting summaries, and in many cases that is where the real work of getting an organization aligned actually occurs.
And here is the part I want to be precise about, because it is the part that connects this back to AI.
A large portion of the most important things that happen at a company happen before the meeting or after the meeting. They happen in the hallway on the way back to the elevator. They happen during the coffee refill. They happen when one person catches another in the parking lot and says the thing they were not going to say in the formal setting. They happen during the dinner after the meeting where the formal agenda has fallen away. They happen when someone walks into your office afterward and closes the door.
None of those things are in the meeting transcript. None of them are captured by the AI note-taker. None of them show up in the summary that gets emailed around afterward. The AI tools are getting genuinely good at capturing what was said during the meeting. They are not capturing the substantial volume of consequential conversation that happens around the meeting, which in many cases is where the real work of getting an organization aligned actually occurs. This is the same shape as the gap between an AI demo and an AI deployment, and it is going to bend operator behavior in the same direction.
I want to be careful with this point, because I have written a lot about how AI is going to compress operational workflows, and I believe that more than I did a year ago. AI note-taking is a useful tool. AI meeting summaries are a useful tool. The fact that I am pointing out something they cannot do is not a critique of the tools. It is a critique of the assumption that the tools cover the full surface area of what a meeting actually is.
The meeting is not only the meeting. The meeting is everything that happens around the meeting, and the relational density between the people who attend it. AI is excellent at capturing the meeting. It is not yet capable of capturing what happens around the meeting, and it cannot create the relational density that makes the around-the-meeting conversations possible in the first place.
That is what I was reminded of this week.
Is remote work bad for company productivity?
No. Remote work has produced real productivity gains and most operational meetings genuinely belong on video. The argument is that some categories of conversation, particularly those involving trust, hard decisions, alignment, conflict, mentorship, or relational dynamics, lose fidelity over video and benefit from being moved to a physical room. The right answer is not full remote or full in person; it is knowing which conversations need which medium.
Remote work made us more efficient. I am not going to argue against that. The ability to pull together a quick meeting with the right people from three different cities in twenty minutes is a productivity multiplier I would never want to give back. The ability to handle a fast operational decision without flying anyone anywhere is a real economic advantage. Most of what runs through video calls every day genuinely belongs on video calls.
But some things do not.
Which conversations need to happen in a physical room?
Conversations where trust is being built, where hard decisions are being made, where a team needs to align on something difficult, where conflict needs to be worked through, where a new hire needs to be integrated into the culture, where a senior person needs to mentor a junior person on something subtle, or where someone needs to push back on the leader without their pushback being processed through the asymmetric power of a remote setup. Those conversations are structurally different in person, not marginally different.
The conversations where trust is being built, where hard decisions are being made, where a team needs to align on something difficult, where conflict needs to be worked through, where a new hire needs to be integrated into the culture, where a senior person needs to mentor a junior person on something subtle, where someone needs to push back on the leader without their pushback being processed through the asymmetric power of a remote setup. Those conversations are different in person, and the difference is not marginal. The difference is structural. The dynamic is the same one I described in Moving at the Speed of Trust: trust is the rate-limiting variable, and the medium you choose either accelerates it or quietly compresses it.
Trust loves proximity. That is what I want operators reading this to take away.
How should leaders decide which meetings need to be in person?
Leaders should ask three questions about their most important conversations. Which of them are still happening over video because of inertia rather than because video is the right medium. Which of them would benefit from being moved to a physical room. Which of them are losing fidelity, alignment, or relational density that the team is not measuring because the team has stopped expecting it to be there. The conversations that surface in those questions are the ones that need a room.
The companies that are figuring out the right mix of remote and in-person work are the ones that will compound advantage over the next decade. The companies that are defaulting to fully remote because it is convenient, or fully in-person because they are nostalgic, are going to underperform the companies that have figured out which conversations need to happen in the room. There is a structural question every leader should be asking right now. Which of our most important conversations are still happening over video because of inertia? Which of them would benefit from being moved to a physical room? Which of them are losing fidelity, alignment, or relational density that we are not measuring because we have stopped expecting it to be there?
For my own team, the answer this week was clear. Some of the conversations we have been running over video for months should have been in person. Some of the things we have been resolving over Slack should have been resolved in someone's office. Some of the alignments we thought we had are actually closer to assumed alignments that have not been pressure-tested in a room.
That is not a Zoom problem. That is not an AI problem. That is a human problem, and the fact that AI tools have made the remote default feel comprehensive has made it easier to forget that the human problem exists.
Here is what I am going to do differently going forward. I am going to be more deliberate about which meetings get a link and which get a room. The link meetings should be the operational ones, the fast-paced ones, the ones where the agenda is straightforward and the relational stakes are low. The room meetings should be the ones where the conversation has emotional weight, where the relational dynamics are complex, where the trust between participants is being built or tested or repaired, where the second-order conversations matter as much as the formal agenda.
That is a bar most operators have stopped applying. The default has become the link. The room is treated as a quaint luxury. I want to put the bar back where it belongs. Not because remote is bad. Because some meetings need a room, not a link.
The AI tools are going to keep getting better at capturing meetings. They will not get better at capturing what happens around meetings, because what happens around meetings is not what an AI is built to capture. The relational substrate of human work is going to remain a human responsibility for as long as humans are doing the work together.
That is what every conversation about the future of meetings, of work, and of teams should actually be about.
Judd Hoffman is CEO and Co-Founder of Ethica AI, building AI-powered tools for real estate transaction workflows.